Demographic and psychological moderators of the relationship between neighborhood cigarette advertising and current smoking in New York City.

BACKGROUND: Tobacco advertising in retailers influences smoking, but little research has examined how this relationship differs among population subgroups. This study merged data on neighborhood cigarette advertising with geocoded survey data to assess the association between advertising prevalence and current smoking among New York City (NYC) residents, and whether demographic and psychological characteristics moderate this relationship. METHODS: Audit data from a stratified, random sample of 796 NYC tobacco retailers generated neighborhood prevalence estimates of cigarette advertising, which were linked with unweighted 2017 NYC Community Health Survey data (n = 7837 adult respondents with residential geocodes). Multilevel regression estimated adjusted odds ratios (aOR) of current smoking by level of neighborhood cigarette advertising (quartiles). Interactions assessed differences in this relationship by demographic characteristics and current depression (analyses conducted in 2019). RESULTS: There was no main effect of advertising on smoking status or significant interactions with demographic variables, but current depression was an effect modifier (p = 0.045). Cigarette advertising was associated with current smoking among those with current depression (p = 0.023), not those without (p = 0.920). Specifically, respondents with depression who resided in neighborhoods in the highest quartile for cigarette advertising prevalence had higher odds of current smoking, compared to those living in the lowest advertising quartile [aOR: 1.72 (1.04, 2.86)]. CONCLUSION: Retail cigarette advertising may serve as an environmental cue to smoke among adults with depression. Efforts to restrict or counteract this practice, such as the development of community-level public health interventions and counter-marketing programs, may particularly benefit those with depression and, perhaps, other mental health disorders.

A comparison of the four healthy days measures (HRQOL-4) with a single measure of self-rated general health in a population-based health survey in New York City.

BACKGROUND: Data on health-related quality of life (HRQOL) can be used to track health disparities, assess the impact of chronic diseases, and predict mortality. The Centers for Disease Control and Prevention's "Healthy Days Measures" (HRQOL-4) assesses four key domains: self-rated general health, physical health, mental health, and activity limitations. The domains are not easily combined to summarize overall HRQOL, and some evidence suggests that self-rated general health may be an adequate proxy indicator for overall HRQOL. This study compares self-rated general health as a solitary measure of HRQOL with two summary indices of the HRQOL-4 as a predictor of adverse health conditions in a representative sample of adult New York City residents. METHODS: The 2017 NYC Social Determinants of Health survey implemented by the New York City Department of Health and Mental Hygiene collected data from a representative sample of New Yorkers (n = 2335) via phone, mail, and web. We compared the information criteria and predictive power of self-rated general health with two alternative summary indices of the HRQOL-4 in predicting self-reported health conditions (hypertension, diabetes, obesity, non-specific psychological distress, and a summary indicator for at least one those four morbidities). RESULTS: Overall, 19.1% (95% CI: 16.9, 21.5) of respondents reported that they had fair or poor general health. Self-rated general health was significantly associated with days of poor physical health, poor mental health, and activity limitations (p < 0.001 for each). While the Akaike and Bayesian information criteria suggested that the summary indices of the HRQOL-4 produced marginally better models for predicting adverse health conditions, self-rated general health had slightly higher predictive power than did the summary indices in all models of physical health outcomes as measured by Tjur's pseudo-R2 and the area under the curve. CONCLUSION: We found very small differences between self-rated general health and the summary indices of the HRQOL-4 in predicting health conditions, suggesting self-rated general health is an appropriate proxy measure of overall HRQOL. Because it can be measured with a single question rather than four, it might be the most simple, efficient, and cost-effective method of summarizing HRQOL in large population-based surveys.